6 BULLETIN 1472, IT. S. DEPARTMENT OF AGRICULTURE 
made at two-hour intervals as to the number of each species of fly 
within the jars. Two observations were made on the day of setting 
out the jars, four each on the second, third, and fourth days of 
exposure, and two observations on the fifth and last day of exposure, 
making 16 observations in all. At the end of each observation period 
the jars were interchanged in position so as to equalize the conditions 
of shade and sunlight as much as possible. In Uvalde the jars were 
set out on the ground in the partial shade of mesquite trees, and 
examined as described above. The results at the two stations, 
Dallas and Uvalde, are similar, and in summarizing the data no 
distinction has been made as to locality. 
In this series of experiments no attempt has been made to deter- 
mine how the meat was rendered unattractive to the flies. It is 
certain, however, that what has been spoken of as repellent action is 
a very complicated matter. It is evident that the meat in these 
tests was protected in several ways by different materials. In some 
cases the protection was largely mechanical, either by covering the 
attractive surface or searing the surface so as to denature the meat 
and stop decomposition; in other cases it was brought about by 
masking the attractive odor of the baits; and in still others it was 
due either to a negative chemotropic response on the part of the fly 
through the sense of smell or an irritation response through the 
respiratory tract or elsewhere. 
METHOD OF COMPUTING RESULTS 
The repellent value of a material is determined by the ratio of the 
number of flies visiting treated meat to the number visiting untreated 
meat. Owing to the great variation in the prevalence of flies from 
week to week, several tests made at different times are necessary to 
accurately gauge the repellent value of any material. In summariz- 
ing these data on repellent action the number of flies of the same 
species visiting all jars treated with the same repellent has been 
used, and the ratio between this number and the number of flies 
visiting a comparable number of untreated or check jars has been 
determined. For example, if the ratios in several tests made at 
different times are 8/119, 23/97, 19/207, these are combined into the 
single ratio 50/423. In this way the observations are weighted 
according to the abundance of flies, as indicated by the number of 
flies visiting the untreated meat. 
The percentage ratio as given is therefore not the percentage of 
repellent efficiency directly, but is the percentage of flies entering 
the treated jars as compared with the number entering the corre- 
sponding checks; that is, a percentage ratio of indicates perfect 
repellent action, 100 shows no effect of the material, and over 100 
indicates that the material is attractive. 
The percentage ratios for the daily periods have not been com- 
puted, but the actual number of flies visiting the jars on each day 
is given. The figures for the first day really represent only one 
afternoon, as the tests were usually begun about midday; and the 
figures for the fifth day usually cover only the forenoon of the last 
day of exposure, as the tests were usually terminated at noon. It is 
believed that some idea of the duration of repellent effect may be 
gained from the comparison of the number of Cochliomyia adults 
